INDEX

Explanations

words related to specific TV shows and movies

oai_token-act-pair · gpt-3.5-turbo

titles of television shows and movies

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 fiber

-0.70

 sanity

-0.69

 contention

-0.69

 affili

-0.68

 credibility

-0.67

 livelihood

-0.67

inventoryQuantity

-0.65

phrine

-0.64

 protective

-0.64

 morale

-0.63

POSITIVE LOGITS

 Awakens

1.12

 Trilogy

1.01

 Bride

0.93

 Incident

0.93

 Below

0.93

 Goes

0.89

 Files

0.88

eers

0.87

 Album

0.87

 Room

0.86

Activations Density 0.180%

words related to specific TV shows and movies

titles of television shows and movies

No Comments

No Known Activations