INDEX

Explanations

instances where someone is attempting or making an effort to do something

oai_token-act-pair · gpt-3.5-turbo

instances of the word "tried."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

head

-0.70

inant

-0.68

scribe

-0.68

Tal

-0.67

cum

-0.66

front

-0.63

cised

-0.63

Trivia

-0.62

effect

-0.60

ificantly

-0.59

POSITIVE LOGITS

 unsuccessfully

1.06

 tried

0.89

anke

0.83

":"/

0.79

 desperately

0.75

 repeatedly

0.73

nir

0.71

 harder

0.71

 frantically

0.70

try

0.69

Activations Density 0.024%

instances where someone is attempting or making an effort to do something

instances of the word "tried."

No Comments

No Known Activations