INDEX

Explanations

news-related terms

oai_token-act-pair · gpt-3.5-turbo

information related to news and updates

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

livious

-0.75

©¶æ¥µ

-0.74

agall

-0.66

gling

-0.65

erer

-0.62

POSE

-0.61

 cumbers

-0.60

 Rite

-0.60

 plurality

-0.59

SpaceEngineers

-0.58

POSITIVE LOGITS

 pertaining

1.02

 from

0.98

 regarding

0.92

 updates

0.89

 relating

0.88

 FROM

0.85

 about

0.85

 subscribe

0.84

 related

0.83

hots

0.82

Activations Density 0.128%

news-related terms

information related to news and updates

No Comments

No Known Activations