INDEX

Explanations

quantitative information related to specific measurements such as amounts or durations

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "approximately" 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 @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

ters

-0.78

woods

-0.73

enders

-0.73

ieu

-0.73

oli

-0.72

Reviewer

-0.72

hered

-0.71

olog

-0.70

lers

-0.66

ny

-0.65

POSITIVE LOGITS

Ĥİ

0.87

 Approximately

0.79

(~

0.78

 midway

0.75

0.74

roximately

0.73

atility

0.73

approximately

0.73

 âĸĪ

0.72

 eighty

0.72

Activations Density 0.018%

quantitative information related to specific measurements such as amounts or durations

occurrences of the word "approximately" in various contexts

No Comments

No Known Activations