INDEX

Explanations

articles (a, an, the)

oai_token-act-pair · gpt-3.5-turbo

articles, specifically the words "a" and "an."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

UD

-0.74

UDP

-0.67

 Provision

-0.63

PID

-0.62

ank

-0.61

 Nemesis

-0.60

 gear

-0.60

 Submission

-0.60

 Sadd

-0.60

 braking

-0.58

POSITIVE LOGITS

ocratic

0.98

emic

0.91

esthetic

0.91

ms

0.90

ria

0.90

alian

0.87

wn

0.87

omorph

0.86

ocracy

0.86

oca

0.83

Activations Density 0.142%

articles (a, an, the)

articles, specifically the words "a" and "an."

No Comments

No Known Activations