INDEX

Explanations

names of specific individuals

oai_token-act-pair · gpt-3.5-turbo

proper nouns, specifically names of individuals

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

tumblr

-0.73

 Marketable

-0.71

 polluted

-0.64

 derailed

-0.63

 coincided

-0.62

 chemicals

-0.62

 coincides

-0.61

sexual

-0.61

 constitutes

-0.61

 engulfed

-0.60

POSITIVE LOGITS

APS

0.72

 Jace

0.71

anton

0.70

UA

0.68

Kis

0.67

 Atkinson

0.67

 Angelo

0.67

cia

0.66

NP

0.65

 Arch

0.64

Activations Density 0.195%

names of specific individuals

proper nouns, specifically names of individuals

No Comments

No Known Activations