INDEX

Explanations

phrases related to location and positional relationships

oai_token-act-pair · gpt-3.5-turbo

references to legal or regulatory concepts and their thresholds

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 ahead

-0.73

isky

-0.71

behind

-0.70

 Cosponsors

-0.62

 differently

-0.62

 wisely

-0.62

wen

-0.62

ONG

-0.61

ãĤ´

-0.61

ãĥĩãĤ£

-0.61

POSITIVE LOGITS

 threshold

0.99

 bounds

0.90

 horizon

0.88

 confines

0.80

 limits

0.79

 boundaries

0.76

ð

0.74

 thresholds

0.74

 scenes

0.73

 level

0.70

Activations Density 0.169%

phrases related to location and positional relationships

references to legal or regulatory concepts and their thresholds

No Comments

No Known Activations