INDEX

Explanations

references to years, particularly the year 2005

oai_token-act-pair · gpt-3.5-turbo

references to the year 2005 and related dates

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

 sympath

-0.63

rx

-0.57

legraph

-0.57

crow

-0.57

 sten

-0.56

isphere

-0.56

 wandered

-0.54

lessly

-0.53

othe

-0.53

Owl

-0.53

POSITIVE LOGITS

3.29

2.80

2.77

2.63

2.54

2.44

2.43

2.41

Activations Density 0.029%

references to years, particularly the year 2005

references to the year 2005 and related dates

No Comments

No Known Activations