INDEX

Explanations

time-related references and durations, such as "years" or "months"

oai_token-act-pair · gpt-3.5-turbo

indicators of duration or periods of time experienced by individuals

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

Elsewhere

-0.69

Optional

-0.69

Alternatively

-0.62

 boost

-0.60

 optional

-0.59

ery

-0.59

rounder

-0.58

 Shutterstock

-0.58

ure

-0.55

tnc

-0.55

POSITIVE LOGITS

 lately

0.99

 hasn

0.94

since

0.94

recent

0.93

 NEVER

0.89

 knows

0.83

 Recently

0.82

 never

0.78

been

0.78

htaking

0.77

Activations Density 0.598%

time-related references and durations, such as "years" or "months"

indicators of duration or periods of time experienced by individuals

No Comments

No Known Activations