INDEX

Explanations

names of specific locations or organizations, particularly those related to crime or investigations

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

 hoops

-0.61

 Pigs

-0.60

»Ĵ

-0.59

 Palest

-0.58

é¾

-0.57

 Lovely

-0.57

 innocence

-0.56

 curve

-0.56

 coping

-0.56

ãĤ¤ãĥĪ

-0.56

POSITIVE LOGITS

aults

0.86

ecast

0.84

ember

0.75

arth

0.74

ault

0.74

fried

0.72

etch

0.71

oard

0.70

wed

0.68

rets

0.68

Activations Density 0.104%

names of specific locations or organizations, particularly those related to crime or investigations

No Comments

No Known Activations