INDEX

Explanations

number-related terms and metioned numbers

oai_token-act-pair · gpt-3.5-turbo

concepts and terminology related to measurement

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

ebted

-0.69

pas

-0.69

rador

-0.69

akespe

-0.67

alus

-0.66

aris

-0.66

venture

-0.65

 brackets

-0.65

atform

-0.64

 Tours

-0.64

POSITIVE LOGITS

Avg

0.69

>.

0.69

XT

0.65

 Telesc

0.64

0.63

 cubic

0.63

asured

0.63

ced

0.61

 Isles

0.58

Population

0.57

Activations Density 0.001%

number-related terms and metioned numbers

concepts and terminology related to measurement

No Comments

No Known Activations