INDEX

Explanations

prepositions indicating relationship or connection between entities or concepts

oai_token-act-pair · gpt-3.5-turbo

phrases that refer to connections or relationships, particularly regarding potential culpability or affiliation

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

accompanied

-0.71

bleacher

-0.59

ortment

-0.58

utenberg

-0.58

Different

-0.57

alian

-0.54

den

-0.52

SAN

-0.52

dependent

-0.52

 withd

-0.51

POSITIVE LOGITS

nor

1.76

 anybody

1.62

any

1.60

 either

1.57

 anything

1.55

 anyone

1.53

ANY

1.36

 slightest

1.36

 anymore

1.35

 whatsoever

1.32

Activations Density 0.659%

prepositions indicating relationship or connection between entities or concepts

phrases that refer to connections or relationships, particularly regarding potential culpability or affiliation

No Comments

No Known Activations