INDEX

Explanations

abbreviations or acronyms related to a specific location or institution

oai_token-act-pair · gpt-3.5-turbo

mentions of "CT" (likely referring to Connecticut) throughout the document

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.4.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

wagen

-1.01

leased

-0.83

mentioned

-0.77

volent

-0.76

tackle

-0.74

olulu

-0.71

bucks

-0.70

 myster

-0.70

conservancy

-0.69

ocene

-0.69

POSITIVE LOGITS

RL

0.96

ARDIS

0.87

 Scan

0.81

arette

0.79

ARGET

0.79

 scan

0.78

urtle

0.78

CT

0.77

ICLE

0.76

arre

0.75

Activations Density 0.010%

abbreviations or acronyms related to a specific location or institution

mentions of "CT" (likely referring to Connecticut) throughout the document

No Comments

No Known Activations

abbreviations or acronyms related to a specific location or institution

mentions of "CT" (likely referring to Connecticut) throughout the document

No Comments

No Known Activations