INDEX

Explanations

phrases indicating continuation, endurance, or lack of decline in various phenomena

oai_token-act-pair · gpt-3.5-turbo

phrases indicating the persistence or continuity of trends and conditions

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

adelphia

-0.77

chens

-0.67

omers

-0.67

 Walters

-0.66

eenth

-0.65

 Centers

-0.64

chool

-0.63

quet

-0.62

ttes

-0.62

 ILCS

-0.62

POSITIVE LOGITS

 truce

0.82

 decay

0.77

 progress

0.75

 improvement

0.74

 aggression

0.74

 differentiation

0.73

 bias

0.72

 validation

0.72

improve

0.72

 signal

0.71

Activations Density 0.069%

phrases indicating continuation, endurance, or lack of decline in various phenomena

phrases indicating the persistence or continuity of trends and conditions

No Comments

No Known Activations