INDEX

Explanations

phrases or words related to actions or statements of assertion

oai_token-act-pair · gpt-3.5-turbo

the word "out" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

wake

-0.74

 ashore

-0.70

avorite

-0.70

avery

-0.68

 scalp

-0.68

 misunder

-0.65

 wallets

-0.65

Syd

-0.62

 submar

-0.61

 oxid

-0.61

POSITIVE LOGITS

atis

0.98

door

0.90

lier

0.89

nels

0.89

dated

0.89

casts

0.85

come

0.84

going

0.83

fits

0.82

fitted

0.82

Activations Density 0.016%

phrases or words related to actions or statements of assertion

the word "out" in various contexts

No Comments

No Known Activations