INDEX

Explanations

media-related terms, particularly related to playback and captioning

oai_token-act-pair · gpt-3.5-turbo

mentions of "Media"

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

bats

-0.72

itudinal

-0.67

SOM

-0.67

ainer

-0.65

aternity

-0.64

 Plum

-0.62

dylib

-0.61

igraph

-0.60

 Xavier

-0.60

tein

-0.60

POSITIVE LOGITS

 playback

1.37

 Coverage

0.96

 outlets

0.91

 enqu

0.89

eval

0.88

 Matters

0.83

 outlet

0.83

 inquiries

0.81

 frenzy

0.77

 mogul

0.77

Activations Density 0.019%

media-related terms, particularly related to playback and captioning

mentions of "Media"

No Comments

No Known Activations