INDEX

Explanations

phrases related to revealing or uncovering something

oai_token-act-pair · gpt-3.5-turbo

references to barriers or obstructions, symbolized by terms like "curtain" and "veil."

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

£ı

-0.96

emort

-0.73

 nomine

-0.71

oyal

-0.70

thritis

-0.69

hab

-0.69

izoph

-0.68

İĭ

-0.68

ciating

-0.67

itaire

-0.66

POSITIVE LOGITS

 curtains

1.11

 curtain

1.08

 veil

0.96

 Veil

0.88

 shroud

0.83

glass

0.76

 clouds

0.72

 Falls

0.72

 draped

0.71

glers

0.70

Activations Density 0.038%

phrases related to revealing or uncovering something

references to barriers or obstructions, symbolized by terms like "curtain" and "veil."

No Comments

No Known Activations