INDEX

Explanations

phrases related to conflicts or disputes

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

instances of words or phrases that indicate significant or exaggerated descriptions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 12-res-jb

Configuration

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

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_post

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

reditary

-0.70

ocious

-0.65

ivism

-0.63

quartered

-0.63

reet

-0.63

criptions

-0.63

worms

-0.61

iotic

-0.59

lings

-0.59

ivist

-0.58

POSITIVE LOGITS

ANCE

0.69

ãĥŃ

0.68

ek

0.66

PS

0.66

LE

0.65

ICE

0.63

ANC

0.62

Ð¾

0.61

ESC

0.61

UME

0.60

Activations Density 0.143%

phrases related to conflicts or disputes

instances of words or phrases that indicate significant or exaggerated descriptions

No Comments

No Known Activations