INDEX

Explanations

information related to research studies, findings, and analyses

oai_token-act-pair · gpt-3.5-turbo

phrases indicating the presence of research findings or claims made by authors

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

 ãĤµãĥ¼ãĥĨãĤ£ãĥ¯ãĥ³

-0.73

 Utility

-0.69

 Playoff

-0.67

irst

-0.66

 Winning

-0.66

ousands

-0.65

 SEAL

-0.64

 Selling

-0.64

 Reporter

-0.62

POSITIVE LOGITS

 besides

1.00

worldly

0.82

ioch

0.81

inois

0.78

 similarly

0.76

than

0.73

 perspectives

0.73

ean

0.73

 confir

0.71

 elig

0.71

Activations Density 0.549%

information related to research studies, findings, and analyses

phrases indicating the presence of research findings or claims made by authors

No Comments

No Known Activations