INDEX

Explanations

time-related phrases, specifically mentioning durations such as "months" or "years"

oai_token-act-pair · gpt-3.5-turbo

references to time periods, specifically those that describe events occurring in the recent past

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Franch

-0.66

kaya

-0.66

aband

-0.65

olas

-0.65

\\\\\\\\

-0.65

/#

-0.65

ensical

-0.62

relative

-0.62

halla

-0.62

cil

-0.61

POSITIVE LOGITS

 gasp

1.04

 month

1.01

 week

1.00

 year

0.98

 ditch

0.98

decade

0.96

ebin

0.91

 night

0.87

 supper

0.84

 Gleaming

0.82

Activations Density 0.063%

time-related phrases, specifically mentioning durations such as "months" or "years"

references to time periods, specifically those that describe events occurring in the recent past

No Comments

No Known Activations