INDEX

Explanations

phrases involving subtle hints or traces of something

oai_token-act-pair · gpt-3.5-turbo

phrases indicating subtle or implied qualities

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

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Merit

-0.88

Reviewer

-0.78

imens

-0.76

erto

-0.75

largest

-0.74

rer

-0.73

 WATCHED

-0.73

riers

-0.72

ļéĨĴ

-0.71

>>\

-0.71

POSITIVE LOGITS

 humor

1.07

 humour

1.01

 irony

0.98

 sweetness

0.96

 cynicism

0.95

 realism

0.93

 brilliance

0.93

 sanity

0.91

 bitterness

0.90

 sadness

0.90

Activations Density 0.205%

phrases involving subtle hints or traces of something

phrases indicating subtle or implied qualities

No Comments

No Known Activations