INDEX

Explanations

words related to selection processes or considerations

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 irony

-0.70

unts

-0.65

nces

-0.63

 exacerbated

-0.62

fortunately

-0.60

 tricked

-0.55

tumblr

-0.54

inis

-0.54

 frustration

-0.53

 Converted

-0.53

POSITIVE LOGITS

"],"

0.74

 next

0.73

OTUS

0.73

 someday

0.71

 candidacy

0.70

 hereafter

0.69

ãĤ¦

0.69

 future

0.68

 NEXT

0.66

Activations Density 1.063%

words related to selection processes or considerations

No Comments

No Known Activations