INDEX

Explanations

words related to the name "Key"

oai_token-act-pair · gpt-3.5-turbo

repeated mentions of "key" related to significant subjects or entities in the document

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

 Expend

-0.67

ATIONAL

-0.66

senal

-0.65

 civilized

-0.64

 disbelief

-0.63

 Mutual

-0.63

 Echoes

-0.60

 Surviv

-0.59

Interstitial

-0.59

 millenn

-0.58

POSITIVE LOGITS

key

1.14

hole

0.99

keys

0.98

*/(

0.97

Key

0.92

holes

0.92

ilant

0.90

Keys

0.89

emort

0.88

pad

0.88

Activations Density 0.007%

words related to the name "Key"

repeated mentions of "key" related to significant subjects or entities in the document

No Comments

No Known Activations