INDEX

Explanations

numerical references to the number eleven

oai_token-act-pair · gpt-3.5-turbo

references to the word "eleven" and its variations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

sburgh

-0.80

 Kenobi

-0.80

lain

-0.72

Ki

-0.69

raints

-0.69

 Keane

-0.69

 Fargo

-0.68

 Papers

-0.67

aldehyde

-0.66

 Hilton

-0.65

POSITIVE LOGITS

venth

1.40

phant

1.26

fter

0.88

oton

0.87

lect

0.84

ven

0.83

LECT

0.81

asure

0.81

ither

0.81

izabeth

0.80

Activations Density 0.031%

numerical references to the number eleven

references to the word "eleven" and its variations

No Comments

No Known Activations