INDEX

Explanations

words related to illumination or visual clarity

oai_token-act-pair · gpt-3.5-turbo

instances of the word "light" 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

å¥³

-0.68

berman

-0.64

 epid

-0.63

accompan

-0.61

BER

-0.59

silver

-0.58

NRS

-0.57

 Soda

-0.57

 à¨

-0.55

VO

-0.55

POSITIVE LOGITS

ight

1.33

mares

0.97

seeing

0.96

forward

0.89

ights

0.85

eous

0.82

rious

0.80

enment

0.79

lights

0.77

rence

0.76

Activations Density 0.006%

words related to illumination or visual clarity

instances of the word "light" in various contexts

No Comments

No Known Activations