INDEX

Explanations

descriptions of physical locations or body parts

oai_token-act-pair · gpt-3.5-turbo

references to the upper body

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

eln

-0.72

 Machina

-0.71

tnc

-0.70

ensable

-0.69

atile

-0.68

YES

-0.67

rx

-0.67

inatory

-0.67

Wikipedia

-0.66

POSITIVE LOGITS

most

1.22

 lobe

0.90

 earners

0.82

jaw

0.80

class

0.80

 limb

0.76

lip

0.76

 bounds

0.76

 priced

0.74

 thigh

0.74

Activations Density 0.039%

descriptions of physical locations or body parts

references to the upper body

No Comments

No Known Activations