INDEX

Explanations

sequences of characters with specific patterns or structure

oai_token-act-pair · gpt-3.5-turbo

scientific notation and references to mathematical or statistical representations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 drink

-0.86

 drinks

-0.84

 Creator

-0.80

 Bible

-0.77

 breakfast

-0.76

 nickname

-0.75

 Lord

-0.75

 bitters

-0.74

 cash

-0.74

 money

-0.74

POSITIVE LOGITS

phase

1.64

surface

1.50

negative

1.50

positive

1.50

response

1.48

phen

1.47

mediated

1.44

specific

1.43

dependent

1.43

termin

1.42

Activations Density 0.133%

sequences of characters with specific patterns or structure

scientific notation and references to mathematical or statistical representations

No Comments

No Known Activations