INDEX

Explanations

phrases related to physical contact or influence

oai_token-act-pair · gpt-3.5-turbo

instances of the word "touched" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

issue

-0.71

adiator

-0.71

asar

-0.70

leys

-0.69

rams

-0.68

orously

-0.67

etheless

-0.65

software

-0.63

 mating

-0.61

ammed

-0.60

POSITIVE LOGITS

 touched

0.97

 touches

0.82

vag

0.77

 GOODMAN

0.76

 fingertips

0.76

 touch

0.73

hou

0.72

stones

0.67

cules

0.67

 toes

0.67

Activations Density 0.008%

phrases related to physical contact or influence

instances of the word "touched" in various contexts

No Comments

No Known Activations