INDEX

Explanations

references to the TV network "FOX" and related content

oai_token-act-pair · gpt-3.5-turbo

mentions of the FOX network

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

adding

-0.77

anks

-0.75

agn

-0.73

oppable

-0.72

idences

-0.72

antics

-0.71

accompan

-0.71

orders

-0.69

anked

-0.69

assisted

-0.69

POSITIVE LOGITS

FOX

1.06

conn

1.00

çīĪ

0.83

 SPORTS

0.81

 Fest

0.76

FOX

0.75

 Broadcast

0.73

TAG

0.73

CAST

0.73

æ©Ł

0.70

Activations Density 0.007%

references to the TV network "FOX" and related content

mentions of the FOX network

No Comments

No Known Activations