INDEX

Explanations

number-related information, particularly focusing on limits or boundaries

oai_token-act-pair · gpt-3.5-turbo

instances of the concept of limits

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

Untitled

-0.68

ochet

-0.66

ette

-0.63

atl

-0.62

Phoenix

-0.60

eye

-0.60

arial

-0.60

 Radiant

-0.59

ÃŃa

-0.59

Pear

-0.58

POSITIVE LOGITS

 limits

4.10

 Limits

3.10

limits

2.59

 limit

2.58

 limitations

2.50

 boundaries

2.26

 bounds

2.14

limit

1.96

 Limit

1.95

 limitation

1.82

Activations Density 0.019%

number-related information, particularly focusing on limits or boundaries

instances of the concept of limits

No Comments

No Known Activations