INDEX

Explanations

phrases related to news articles or official statements

oai_token-act-pair · gpt-3.5-turbo

names of individuals or entities involved in a legal or official context

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

ModLoader

-0.53

 clubhouse

-0.52

tumblr

-0.51

 robbers

-0.49

 fodder

-0.47

 symbolic

-0.45

 ferment

-0.45

 styl

-0.45

 gearing

-0.45

 tyre

-0.45

POSITIVE LOGITS

(@

0.76

ovich

0.73

mann

0.68

ansky

0.67

oulos

0.67

inski

0.66

zinski

0.64

auer

0.63

anan

0.62

uria

0.61

Activations Density 0.535%

phrases related to news articles or official statements

names of individuals or entities involved in a legal or official context

No Comments

No Known Activations