INDEX

Explanations

references to specific time durations, particularly focusing on weeks

oai_token-act-pair · gpt-3.5-turbo

references to specific durations of weeks 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

inav

-0.87

cellence

-0.68

backer

-0.68

acted

-0.64

grab

-0.64

netic

-0.63

Reviewer

-0.63

ech

-0.63

Dub

-0.61

emort

-0.61

POSITIVE LOGITS

days

1.04

ago

0.96

 gestation

0.87

DAY

0.86

iversary

0.82

Ago

0.79

night

0.77

oldown

0.71

cakes

0.70

uary

0.69

Activations Density 0.046%

references to specific time durations, particularly focusing on weeks

references to specific durations of weeks in various contexts

No Comments

No Known Activations