INDEX

Explanations

names or pronouns indicating a person that is being spoken to

oai_token-act-pair · gpt-3.5-turbo

references to conversations or the act of talking

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

©¶æ¥µ

-0.81

pite

-0.78

amins

-0.73

ŃĶ

-0.72

irez

-0.70

ģĸ

-0.69

works

-0.67

gross

-0.65

({

-0.65

 Goods

-0.64

POSITIVE LOGITS

him

0.90

us

0.76

wered

0.72

 them

0.71

me

0.70

 passers

0.69

 reporters

0.68

 someone

0.68

 strangers

0.68

 voic

0.67

Activations Density 0.054%

names or pronouns indicating a person that is being spoken to

references to conversations or the act of talking

No Comments

No Known Activations