INDEX

Explanations

phrases related to conflicts of interest

oai_token-act-pair · gpt-3.5-turbo

phrases related to conflicts of interest

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

ãĥ¥

-0.80

aper

-0.76

OCK

-0.73

Cru

-0.73

jong

-0.73

ISS

-0.70

pload

-0.70

nington

-0.70

NAS

-0.69

é¾įåĸļå£«

-0.69

POSITIVE LOGITS

 interests

0.84

arial

0.83

 liability

0.76

 materially

0.76

 entities

0.76

 Ethics

0.75

 valves

0.72

 considerations

0.72

awaru

0.72

 liabilities

0.71

Activations Density 0.019%

phrases related to conflicts of interest

phrases related to conflicts of interest

No Comments

No Known Activations