INDEX

Explanations

instances of the word "drove" in various contexts

oai_token-act-pair · gpt-3.5-turbo

instances of the word "drove" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

 Seym

-0.73

aido

-0.71

icipated

-0.65

eon

-0.65

Lum

-0.65

ANN

-0.65

assium

-0.64

achu

-0.64

çĦ

-0.64

ency

-0.63

POSITIVE LOGITS

 driving

0.92

wheel

0.82

driving

0.81

 drives

0.80

 drove

0.78

driver

0.77

HAM

0.76

 Driving

0.75

 slic

0.72

 drunk

0.69

Activations Density 0.009%

instances of the word "drove" in various contexts

instances of the word "drove" in various contexts

No Comments

No Known Activations