INDEX

Explanations

phrases related to technology and power

oai_token-act-pair · gpt-3.5-turbo

terms related to official or government-related documents and statements

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

itta

-0.79

asso

-0.76

adier

-0.71

atel

-0.69

KEY

-0.67

BIT

-0.65

oother

-0.65

rolog

-0.65

codes

-0.64

oter

-0.61

POSITIVE LOGITS

theless

0.96

tenance

0.86

terday

0.85

lihood

0.79

nesses

0.78

foundland

0.78

quartered

0.75

etheless

0.75

looking

0.70

CLASSIFIED

0.70

Activations Density 0.119%

phrases related to technology and power

terms related to official or government-related documents and statements

No Comments

No Known Activations