INDEX

Explanations

proper nouns indicating specific locations or neighborhoods

oai_token-act-pair · gpt-3.5-turbo

possessive forms indicating locations or attributes associated with various places

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

itcher

-0.81

arians

-0.75

ggie

-0.74

CVE

-0.73

JECT

-0.70

Reviewer

-0.70

ISO

-0.70

hari

-0.69

ENN

-0.68

Pak

-0.67

POSITIVE LOGITS

own

1.10

 infancy

0.95

 newest

0.95

 finest

0.88

 hometown

0.88

 inaugural

0.86

 behalf

0.86

 infamous

0.86

 periphery

0.86

 imagination

0.85

Activations Density 0.135%

proper nouns indicating specific locations or neighborhoods

possessive forms indicating locations or attributes associated with various places

No Comments

No Known Activations