INDEX

Explanations

instances where the word "participate" is mentioned

oai_token-act-pair · gpt-3.5-turbo

instances of the word "participate."

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

tight

-0.77

forward

-0.76

cut

-0.70

brow

-0.70

eric

-0.65

south

-0.65

deal

-0.64

cold

-0.63

flu

-0.62

separ

-0.62

POSITIVE LOGITS

 participates

1.03

 participate

1.02

 participation

0.94

 Participation

0.91

ivity

0.91

 guiActiveUn

0.90

 participated

0.85

udo

0.84

ADRA

0.80

 participating

0.79

Activations Density 0.013%

instances where the word "participate" is mentioned

instances of the word "participate."

No Comments

No Known Activations