INDEX

Explanations

words related to the action of draining

oai_token-act-pair · gpt-3.5-turbo

references to the word "drain" and its variations in different contexts

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ordan

-0.80

 Flavoring

-0.78

*/(

-0.73

ourke

-0.70

eer

-0.69

fman

-0.67

eering

-0.66

MET

-0.66

 Legal

-0.66

PRES

-0.65

POSITIVE LOGITS

 drain

1.32

 drains

1.14

 drained

0.96

 draining

0.95

 Drain

0.93

pipe

0.85

 irrigation

0.81

 condu

0.73

wright

0.73

 siph

0.73

Activations Density 0.010%

words related to the action of draining

references to the word "drain" and its variations in different contexts

No Comments

No Known Activations