INDEX

Explanations

terms related to adjustments and settings

oai_token-act-pair · gpt-3.5-turbo

references to 'cleats'

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

verified

-0.73

Operation

-0.69

 Residents

-0.69

bah

-0.66

 EVENT

-0.65

registered

-0.64

pmwiki

-0.64

XM

-0.64

tests

-0.64

ishi

-0.63

POSITIVE LOGITS

cle

4.10

Cle

1.18

 pard

1.02

Cle

0.93

 hars

0.88

 pardon

0.86

 widest

0.83

le

0.80

 bras

0.80

 fullest

0.79

Activations Density 0.015%

terms related to adjustments and settings

references to 'cleats'

No Comments

No Known Activations