INDEX

Explanations

instances where the token "Nat" is present

oai_token-act-pair · gpt-3.5-turbo

references to the name "Nat."

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

BACK

-0.75

âĹ¼

-0.72

wordpress

-0.70

Ħ¢

-0.68

AGES

-0.67

ngth

-0.66

ragon

-0.65

STON

-0.65

deen

-0.64

initialized

-0.61

POSITIVE LOGITS

itionally

1.02

ritional

1.00

uggets

0.98

omas

0.94

ificent

0.93

ual

0.92

ulia

0.90

rition

0.89

Ðµ

0.89

ured

0.88

Activations Density 0.007%

instances where the token "Nat" is present

references to the name "Nat."

No Comments

No Known Activations