INDEX

Explanations

television network names and programs

oai_token-act-pair · gpt-3.5-turbo

references to television networks and broadcasting schedules

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

crete

-0.80

 mitigation

-0.79

aez

-0.78

potion

-0.77

 geometry

-0.72

ysics

-0.71

objects

-0.70

ertodd

-0.69

auld

-0.69

Materials

-0.69

POSITIVE LOGITS

 broadcasts

1.53

levision

1.49

 airing

1.48

TV

1.48

 television

1.47

 viewership

1.46

HBO

1.45

 broadcast

1.45

 broadcasting

1.44

CBS

1.43

Activations Density 0.932%

television network names and programs

references to television networks and broadcasting schedules

No Comments

No Known Activations