INDEX

Explanations

references to reading or consuming content

oai_token-act-pair · gpt-3.5-turbo

references to reading and written material

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

aires

-0.79

akable

-0.74

rio

-0.67

ichick

-0.66

XL

-0.64

assador

-0.61

©¶æ¥µ

-0.61

 Regist

-0.60

ãĥł

-0.60

ONSORED

-0.60

POSITIVE LOGITS

 aloud

1.39

 excerpts

0.95

DragonMagazine

0.91

 Digest

0.89

 passages

0.88

 transcript

0.87

 newspapers

0.86

 journals

0.81

 articles

0.81

 transcripts

0.80

Activations Density 0.141%

references to reading or consuming content

references to reading and written material

No Comments

No Known Activations