INDEX

Explanations

phrases related to looking at or acknowledging someone or something

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "at" and related phrases indicating location or direction

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

interstitial

-0.87

ancial

-0.80

ccording

-0.78

artifacts

-0.74

exempt

-0.74

ONSORED

-0.73

umerous

-0.70

 exting

-0.70

quartered

-0.69

endiary

-0.69

POSITIVE LOGITS

him

1.52

me

1.49

us

1.31

him

1.15

 herself

1.12

 himself

1.11

Him

1.09

HIM

1.07

ya

1.07

 them

0.97

Activations Density 0.336%

phrases related to looking at or acknowledging someone or something

occurrences of the word "at" and related phrases indicating location or direction

No Comments

No Known Activations