INDEX

Explanations

negative adjectives

oai_token-act-pair · gpt-3.5-turbo

negative sentiments associated with the word "awful."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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£«

-0.83

coni

-0.82

$$$$

-0.82

KER

-0.76

xtap

-0.73

riott

-0.73

VO

-0.70

externalActionCode

-0.70

FBI

-0.69

ARCH

-0.68

POSITIVE LOGITS

ness

1.10

nesses

0.91

entimes

0.86

 traged

0.83

headed

0.78

lot

0.77

alia

0.75

 blob

0.74

pse

0.73

 nodd

0.72

Activations Density 0.027%

negative adjectives

negative sentiments associated with the word "awful."

No Comments

No Known Activations