INDEX

Explanations

terms related to background information or background checks

oai_token-act-pair · gpt-3.5-turbo

references to background information or background checks

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

erion

-0.73

 ãĤµãĥ¼ãĥĨãĤ£ãĥ¯ãĥ³

-0.72

nard

-0.71

aden

-0.71

emic

-0.69

ap

-0.68

ority

-0.68

orean

-0.68

abase

-0.67

apest

-0.66

POSITIVE LOGITS

 noise

1.05

 checks

1.00

 Checks

0.87

spection

0.83

 check

0.79

 noises

0.78

 radiation

0.78

 information

0.76

 music

0.76

 info

0.74

Activations Density 0.054%

terms related to background information or background checks

references to background information or background checks

No Comments

No Known Activations