INDEX

Explanations

instances of feedback and related terms within text passages

oai_token-act-pair · gpt-3.5-turbo

references to user feedback in various contexts

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

wikipedia

-0.80

 Noct

-0.69

amina

-0.67

adem

-0.63

Dial

-0.63

neys

-0.63

chin

-0.61

ffe

-0.61

 Carey

-0.61

esm

-0.61

POSITIVE LOGITS

 loops

1.15

 loop

0.99

 feedback

0.96

 Feedback

0.88

 sessions

0.83

ollow

0.74

 testers

0.73

0.72

 suggestions

0.70

 constructive

0.69

Activations Density 0.050%

instances of feedback and related terms within text passages

references to user feedback in various contexts

No Comments

No Known Activations