INDEX

Explanations

behavior-related words and phrases

oai_token-act-pair · gpt-3.5-turbo

references to behavior and its various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

endiary

-0.75

mand

-0.72

inite

-0.71

anmar

-0.69

racted

-0.67

sonian

-0.67

enegger

-0.67

inka

-0.67

vu

-0.66

ondo

-0.65

POSITIVE LOGITS

 modification

1.06

 behaviors

1.05

 behavior

1.01

avior

0.97

 behaviours

0.97

aviour

0.96

uation

0.95

 patterns

0.95

 behavi

0.92

uate

0.92

Activations Density 0.049%

behavior-related words and phrases

references to behavior and its various contexts

No Comments

No Known Activations