INDEX

Explanations

phrases related to news or significant occurrences

oai_token-act-pair · gpt-3.5-turbo

references to incidents or occurrences frequently mentioned in a narrative

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

emouth

-0.90

adders

-0.77

azy

-0.76

itte

-0.76

asonry

-0.75

igsaw

-0.75

hov

-0.75

ularity

-0.74

tsky

-0.74

guided

-0.74

POSITIVE LOGITS

 incident

1.06

 incidents

0.92

 involving

0.83

 Incident

0.80

uates

0.79

 occurrence

0.74

uality

0.74

uation

0.72

 rapes

0.70

Dialogue

0.68

Activations Density 0.019%

phrases related to news or significant occurrences

references to incidents or occurrences frequently mentioned in a narrative

No Comments

No Known Activations