INDEX

Explanations

statements describing or categorizing different types of things or behavior

oai_token-act-pair · gpt-3.5-turbo

phrases that describe types or kinds of behavior or incidents

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

stice

-0.60

gov

-0.60

than

-0.59

selves

-0.56

zig

-0.56

ten

-0.56

ened

-0.55

asonic

-0.55

 connection

-0.55

 Accessed

-0.55

POSITIVE LOGITS

yip

0.91

uyomi

0.80

 ordinarily

0.77

 nightmares

0.76

ometimes

0.75

 normally

0.73

 attracts

0.73

 dictators

0.71

uman

0.70

abl

0.69

Activations Density 0.202%

statements describing or categorizing different types of things or behavior

phrases that describe types or kinds of behavior or incidents

No Comments

No Known Activations