INDEX

Explanations

phrases related to results or consequences

oai_token-act-pair · gpt-3.5-turbo

references to outcomes in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

cer

-0.82

nan

-0.77

elong

-0.71

ondo

-0.70

vor

-0.70

ju

-0.69

ker

-0.69

bia

-0.69

entin

-0.68

azines

-0.68

POSITIVE LOGITS

 outcome

1.38

 outcomes

1.13

bringer

0.76

 Result

0.76

 result

0.74

icter

0.71

DragonMagazine

0.70

ilater

0.69

 bystanders

0.68

 winner

0.68

Activations Density 0.008%

phrases related to results or consequences

references to outcomes in various contexts

No Comments

No Known Activations