INDEX

Explanations

mentions of locations or items being positioned next to each other

oai_token-act-pair · gpt-3.5-turbo

phrases indicating spatial relationships or proximity

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ivities

-0.75

ãĤ¢ãĥ«

-0.71

tera

-0.68

TERN

-0.68

appropriate

-0.68

nesota

-0.66

xml

-0.66

iasm

-0.65

omnia

-0.64

?????-?????-

-0.63

POSITIVE LOGITS

 where

1.03

 hers

0.99

 ours

0.91

 them

0.90

 yours

0.89

him

0.87

 parked

0.85

 mine

0.82

us

0.82

 theirs

0.80

Activations Density 0.171%

mentions of locations or items being positioned next to each other

phrases indicating spatial relationships or proximity

No Comments

No Known Activations