INDEX

Explanations

abbreviations or acronyms ending in 'ls'

oai_token-act-pair · gpt-3.5-turbo

occurrences of the phrase "url" and its variations

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

GI

-0.62

 Canad

-0.62

 benefiting

-0.62

 benef

-0.61

 Ward

-0.60

 Comic

-0.59

 inaccur

-0.58

 improving

-0.58

 correction

-0.58

 improvement

-0.57

POSITIVE LOGITS

ls

4.50

LS

1.73

lv

1.63

ln

1.58

ns

1.51

lr

1.49

lt

1.41

lers

1.41

lb

1.41

lf

1.35

Activations Density 0.008%

abbreviations or acronyms ending in 'ls'

occurrences of the phrase "url" and its variations

No Comments

No Known Activations