INDEX

Explanations

personal names or proper nouns, possibly from various languages

oai_token-act-pair · gpt-3.5-turbo

words related to existence or presence

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

RAW

-0.76

clinton

-0.73

SPONSORED

-0.69

 Flavoring

-0.66

Defense

-0.66

warming

-0.65

agonist

-0.64

Cold

-0.63

puff

-0.63

CEO

-0.62

POSITIVE LOGITS

 alle

0.98

qui

0.89

mi

0.85

cknowled

0.84

ande

0.83

ja

0.83

 thou

0.82

 este

0.82

si

0.82

que

0.81

Activations Density 0.132%

personal names or proper nouns, possibly from various languages

words related to existence or presence

No Comments

No Known Activations