INDEX

Explanations

These activations suggest that the neuron is looking for names of specific places, people, and entities

oai_token-act-pair · gpt-3.5-turbo

proper nouns or names associated with people and places

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

 defect

-0.70

ufact

-0.69

ucl

-0.67

shire

-0.64

etheus

-0.57

 imperative

-0.56

 narrated

-0.56

ensical

-0.55

arcity

-0.54

 cour

-0.54

POSITIVE LOGITS

hiba

0.80

wagen

0.79

agi

0.74

Mods

0.73

scl

0.70

hei

0.69

kat

0.68

hesda

0.64

apons

0.63

Container

0.62

Activations Density 1.157%

These activations suggest that the neuron is looking for names of specific places, people, and entities

proper nouns or names associated with people and places

No Comments

No Known Activations