INDEX

Explanations

United States Senators

oai_token-act-pair · gpt-3.5-turbo

references to senators

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

tru

-0.69

 unmarked

-0.67

 hitch

-0.64

ATHER

-0.64

ORN

-0.63

 towed

-0.61

mania

-0.59

 misunderstanding

-0.59

footed

-0.58

 Bearing

-0.58

POSITIVE LOGITS

iors

1.54

pai

1.39

seless

1.26

egal

1.26

escent

1.22

eca

1.19

esis

1.01

IOR

0.98

esse

0.98

iture

0.94

Activations Density 0.026%

United States Senators

references to senators

No Comments

No Known Activations