INDEX

Explanations

proper nouns related to locations such as cities and companies

oai_token-act-pair · gpt-3.5-turbo

references to specific locations and names

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

 partName

-0.81

eper

-0.81

usp

-0.78

eva

-0.76

leck

-0.75

akening

-0.75

ipl

-0.74

auri

-0.73

ech

-0.70

letal

-0.69

POSITIVE LOGITS

 Newport

1.26

 Beach

0.90

 SEAL

0.89

 Huntington

0.79

 Coast

0.78

é»Ĵ

0.75

raits

0.74

ãĥ¼ãĥĨãĤ£

0.73

ogue

0.72

 Ranch

0.72

Activations Density 0.014%

proper nouns related to locations such as cities and companies

references to specific locations and names

No Comments

No Known Activations