INDEX

Explanations

phrases related to self-related concepts or actions

oai_token-act-pair · gpt-3.5-turbo

references to self-identity or self-related concepts

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

•henryjames

No Comments

Negative Logits

IUM

-0.75

ICAN

-0.71

nis

-0.69

ONT

-0.67

pheus

-0.62

ondo

-0.62

 Nights

-0.61

dayName

-0.61

andum

-0.60

oS

-0.60

POSITIVE LOGITS

lessly

0.97

same

0.95

 esteem

0.94

destruct

0.93

 destruct

0.93

ridges

0.90

 explanatory

0.88

 proclaimed

0.86

less

0.83

Activations Density 0.032%

phrases related to self-related concepts or actions

references to self-identity or self-related concepts

No Comments

No Known Activations