INDEX

Explanations

proper nouns related to companies or institutions

oai_token-act-pair · gpt-3.5-turbo

proper nouns related to well-known companies and entities

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

onen

-0.73

agonists

-0.69

udic

-0.68

cised

-0.68

 fuzz

-0.68

)].

-0.62

 downstream

-0.62

ecd

-0.61

.?

-0.61

enforcement

-0.60

POSITIVE LOGITS

 Forever

0.76

 Orig

0.75

Care

0.74

 Entertainment

0.74

 Plaza

0.74

 Badge

0.73

 International

0.73

 Expedition

0.72

 Personal

0.72

 Scion

0.72

Activations Density 0.386%

proper nouns related to companies or institutions

proper nouns related to well-known companies and entities

No Comments

No Known Activations