INDEX

Explanations

articles or publications mentioned along with the word "Article"

oai_token-act-pair · gpt-3.5-turbo

references to authorship and publication details

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

auga

-0.66

 awoken

-0.65

zers

-0.64

htar

-0.63

inh

-0.62

untarily

-0.62

 chimpanzees

-0.62

 tranquil

-0.61

icent

-0.61

 ceilings

-0.60

POSITIVE LOGITS

govtrack

0.84

 headlined

0.80

 titled

0.76

 Column

0.72

çīĪ

0.69

 authored

0.68

 headlines

0.66

 editor

0.65

Contents

0.65

comment

0.64

Activations Density 0.311%

articles or publications mentioned along with the word "Article"

references to authorship and publication details

No Comments

No Known Activations