INDEX

Explanations

time-related terms like days, weeks, months, or years

oai_token-act-pair · gpt-3.5-turbo

references to time intervals, specifically durations like weeks and months

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

enburg

-0.70

acus

-0.66

jac

-0.64

inite

-0.63

 Encyclopedia

-0.62

ghazi

-0.61

ocaly

-0.61

ãĤ¢ãĥ«

-0.56

LM

-0.56

govern

-0.55

POSITIVE LOGITS

ago

2.05

Ago

1.34

 back

1.27

 beforehand

1.06

 prior

1.06

 before

0.99

 earlier

0.98

 later

0.96

 gone

0.93

 elapsed

0.92

Activations Density 0.080%

time-related terms like days, weeks, months, or years

references to time intervals, specifically durations like weeks and months

No Comments

No Known Activations