INDEX

Explanations

mentions of the word "Py" followed by a number

oai_token-act-pair · gpt-3.5-turbo

references to a specific product named Pyra

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

nesota

-0.76

DERR

-0.76

 Rockefeller

-0.72

ously

-0.71

 Hearts

-0.71

ional

-0.71

worthiness

-0.70

hips

-0.69

PUT

-0.67

iveness

-0.67

POSITIVE LOGITS

ramids

1.15

Py

0.98

Py

0.97

ongyang

0.96

rote

0.89

romancer

0.86

ramid

0.83

gm

0.82

oleon

0.77

atom

0.76

Activations Density 0.011%

mentions of the word "Py" followed by a number

references to a specific product named Pyra

No Comments

No Known Activations