INDEX

Explanations

TV networks and specific TV shows

oai_token-act-pair · gpt-3.5-turbo

mentions of television networks and streaming services

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ysis

-0.76

 Hast

-0.68

 Adin

-0.65

bitcoin

-0.64

DPR

-0.63

ertodd

-0.63

 subsistence

-0.61

xit

-0.61

 Bitcoin

-0.60

 drilled

-0.60

POSITIVE LOGITS

 Studios

0.95

ovies

0.93

 Films

0.89

 blockbuster

0.86

levision

0.85

 sequels

0.83

 exec

0.79

 Classics

0.79

 adaptation

0.78

 Television

0.78

Activations Density 0.206%

TV networks and specific TV shows

mentions of television networks and streaming services

No Comments

No Known Activations