INDEX

Explanations

expressions of time duration

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate durations of time, particularly years and months

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

cause

-0.69

Mech

-0.68

 helicop

-0.68

Case

-0.67

Fac

-0.67

Pri

-0.65

ORD

-0.63

wealth

-0.62

Saturday

-0.61

ice

-0.60

POSITIVE LOGITS

 hiatus

0.96

ynthesis

0.84

dding

0.83

 slog

0.77

 reckoning

0.73

 exhausting

0.72

 absence

0.70

 exhaustive

0.69

 deliber

0.69

cation

0.68

Activations Density 0.077%

expressions of time duration

phrases that indicate durations of time, particularly years and months

No Comments

No Known Activations