INDEX

Explanations

mentions of categories or groups with a hierarchical or status-related connotation

oai_token-act-pair · gpt-3.5-turbo

references to "under" phrases indicating a ranking or classification

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 Retrieved

-0.62

 goodbye

-0.61

 onward

-0.60

 Osaka

-0.60

 Begins

-0.60

aukee

-0.59

 supra

-0.58

 farewell

-0.58

rob

-0.58

 eternity

-0.57

POSITIVE LOGITS

dogs

1.33

whelming

1.31

developed

1.25

represented

1.24

stated

1.23

study

1.19

performing

1.19

priv

1.16

graduate

1.16

powered

1.16

Activations Density 0.042%

mentions of categories or groups with a hierarchical or status-related connotation

references to "under" phrases indicating a ranking or classification

No Comments

No Known Activations