INDEX

Explanations

dates and titles

oai_token-act-pair · gpt-3.5-turbo

prominent references to specific events or entities in a narrative 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

raints

-0.85

aeda

-0.79

 disadvant

-0.78

luster

-0.78

farious

-0.75

orage

-0.74

olicy

-0.74

obbies

-0.73

Downloadha

-0.72

inki

-0.71

POSITIVE LOGITS

cknowled

0.67

KH

0.61

 Copyright

0.60

 COURT

0.60

Emb

0.58

 translator

0.58

 Photograph

0.58

 transl

0.57

 Rain

0.56

 Opinion

0.55

Activations Density 0.254%

dates and titles

prominent references to specific events or entities in a narrative context

No Comments

No Known Activations