INDEX

Explanations

phrases related to news, updates, and events

oai_token-act-pair · gpt-3.5-turbo

contexts related to forecasts and predictions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 ideally

-0.81

 violate

-0.76

 thereof

-0.76

 violates

-0.76

afety

-0.75

nown

-0.72

Sov

-0.72

 violated

-0.72

 furthermore

-0.72

 beforehand

-0.71

POSITIVE LOGITS

1.01

0.96

0.85

 offseason

0.82

 Charlottesville

0.79

0.77

 Atlanta

0.76

 Wilmington

0.75

Activations Density 0.482%

phrases related to news, updates, and events

contexts related to forecasts and predictions

No Comments

No Known Activations