INDEX

Explanations

specific locations or properties, particularly related to real estate or political figures

oai_token-act-pair · gpt-3.5-turbo

locations and organizations mentioned in the text

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

ians

-0.82

Charges

-0.73

ting

-0.72

wave

-0.71

idas

-0.70

ious

-0.69

 Holt

-0.69

erenn

-0.67

 Zerg

-0.66

 Geoffrey

-0.64

POSITIVE LOGITS

creen

0.92

 Alam

0.87

userc

0.85

 clubhouse

0.85

bara

0.82

Lago

0.81

iferation

0.78

ignt

0.77

psey

0.75

 referen

0.74

Activations Density 0.022%

specific locations or properties, particularly related to real estate or political figures

locations and organizations mentioned in the text

No Comments

No Known Activations