INDEX

Explanations

phrases related to self-identity and introspection

oai_token-act-pair · gpt-3.5-turbo

instances of a specific character or symbol

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

ammy

-0.70

 Gutenberg

-0.67

 fortun

-0.66

 fragmentation

-0.63

 mathemat

-0.63

 pressures

-0.61

 disadvant

-0.59

 Palestin

-0.58

 fertil

-0.57

 intercepted

-0.54

POSITIVE LOGITS

ï¸ı

0.88

ï¸

0.77

sure

0.71

İ

0.70

Balt

0.70

hood

0.67

----------------------------------------------------------------

0.66

hall

0.65

âģ

0.64

Â£

0.64

Activations Density 0.283%

phrases related to self-identity and introspection

instances of a specific character or symbol

No Comments

No Known Activations