INDEX

Explanations

phrases related to news articles and reports

oai_token-act-pair · gpt-3.5-turbo

the presence of empty or non-informative text segments

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

 manif

-0.69

onest

-0.68

cup

-0.67

schild

-0.66

hower

-0.65

rouse

-0.65

forth

-0.65

obbies

-0.65

ccess

-0.64

inse

-0.64

POSITIVE LOGITS

CLUS

0.78

Gat

0.77

actionDate

0.75

 Licensed

0.72

 Hide

0.70

 Latest

0.67

 Wrestling

0.67

UTERS

0.67

.)

0.65

 Reporter

0.65

Activations Density 0.096%

phrases related to news articles and reports

the presence of empty or non-informative text segments

No Comments

No Known Activations