INDEX

Explanations

events or occurrences related to someone's first appearances or achievements

oai_token-act-pair · gpt-3.5-turbo

references to first-time events and significant appearances

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

ruciating

-0.66

aples

-0.65

 preferred

-0.64

 abound

-0.63

 Defin

-0.59

 favorites

-0.58

ometimes

-0.55

ijn

-0.54

ictive

-0.54

itable

-0.54

POSITIVE LOGITS

 since

1.29

 EVER

1.29

 ever

1.20

since

0.93

 Ever

0.91

iversary

0.90

 appearance

0.89

ever

0.86

 Appearance

0.83

 anniversary

0.83

Activations Density 0.205%

events or occurrences related to someone's first appearances or achievements

references to first-time events and significant appearances

No Comments

No Known Activations