INDEX

Explanations

proper nouns or phrases related to intelligence agencies or investigative journalism

oai_token-act-pair · gpt-3.5-turbo

terms related to repetition and conditions in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

hurst

-0.62

 Phoenix

-0.61

 shoulders

-0.58

 Unch

-0.57

 highlights

-0.56

 Pearl

-0.56

 snake

-0.55

 Beaver

-0.55

 Irving

-0.54

 Mermaid

-0.54

POSITIVE LOGITS

idate

0.97

ensable

0.94

igent

0.94

idem

0.91

ACTION

0.90

antage

0.88

inence

0.87

perate

0.87

acies

0.86

usive

0.86

Activations Density 0.062%

proper nouns or phrases related to intelligence agencies or investigative journalism

terms related to repetition and conditions in various contexts

No Comments

No Known Activations