INDEX

Explanations

mentions of specific symbols or formatting elements within text documents

oai_token-act-pair · gpt-3.5-turbo

structural or statistical patterns in research data

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 neighb

-0.78

 swoop

-0.76

 occas

-0.72

 hopping

-0.69

 replacements

-0.67

 whine

-0.67

 wiser

-0.66

 whining

-0.64

 whistle

-0.63

 snap

-0.63

POSITIVE LOGITS

ÂłÂłÂłÂłÂłÂłÂłÂł

1.20

ÂłÂłÂłÂł

1.12

Figure

1.09

ÂłÂłÂł

1.07

posted

1.03

ÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂł

1.03

METHOD

1.02

Nonetheless

1.00

Methods

1.00

âĵĺ

0.98

Activations Density 0.308%

mentions of specific symbols or formatting elements within text documents

structural or statistical patterns in research data

No Comments

No Known Activations