INDEX

Explanations

IDs

oai_token-act-pair · gpt-3.5-turbo

identifiers or codes associated with entities or subjects

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

andise

-1.02

 Seym

-0.73

bilt

-0.72

 silence

-0.67

Nun

-0.64

ruciating

-0.63

ndum

-0.62

orsi

-0.62

 indo

-0.59

 Papua

-0.58

POSITIVE LOGITS

irect

1.15

irection

1.03

iots

1.01

ocument

1.00

imensional

0.98

escription

0.92

DEN

0.90

ouble

0.89

itor

0.87

ividual

0.86

Activations Density 0.015%

IDs

identifiers or codes associated with entities or subjects

No Comments

No Known Activations