INDEX

Explanations

references to specific time periods or durations

oai_token-act-pair · gpt-3.5-turbo

references to time periods or temporal markers

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 lieutenant

-0.69

 stumbling

-0.67

 sight

-0.66

 nest

-0.65

hog

-0.64

 feral

-0.64

 stake

-0.62

 sear

-0.62

 poker

-0.61

 safety

-0.60

POSITIVE LOGITS

 Period

4.03

period

2.01

 period

1.31

 periods

1.21

 Procedures

1.20

Era

1.17

 Dates

1.04

 Cycle

0.95

 Years

0.95

 Parables

0.95

Activations Density 0.015%

references to specific time periods or durations

references to time periods or temporal markers

No Comments

No Known Activations