INDEX

Explanations

phrases related to legal and political matters

oai_token-act-pair · gpt-3.5-turbo

appeals for public assistance or information

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

defense

-0.94

avored

-0.93

colored

-0.90

 favors

-0.89

 realized

-0.89

 utilized

-0.89

 honored

-0.89

imize

-0.88

 unfavorable

-0.87

 favoring

-0.87

POSITIVE LOGITS

 Â£

1.84

 Yorkshire

1.59

 Ukip

1.59

 UKIP

1.57

NHS

1.57

Scotland

1.55

 Scotland

1.55

 Labour

1.54

 Cardiff

1.52

 Edinburgh

1.51

Activations Density 3.120%

phrases related to legal and political matters

appeals for public assistance or information

No Comments

No Known Activations