INDEX

Explanations

references to watching videos or broadcasts

oai_token-act-pair · gpt-3.5-turbo

mentions of video content

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

ctrl

-0.91

interstitial

-0.72

ascal

-0.66

 misunderstanding

-0.66

cffffcc

-0.65

ACC

-0.64

phi

-0.64

rhy

-0.63

ãĤ¨ãĥ«

-0.62

RET

-0.61

POSITIVE LOGITS

 Watching

1.19

tower

1.19

dog

1.09

dogs

0.98

Watch

0.94

 Watch

0.93

ing

0.84

 Dogs

0.84

 WATCH

0.83

watch

0.83

Activations Density 0.015%

references to watching videos or broadcasts

mentions of video content

No Comments

No Known Activations