INDEX

Explanations

references to specific car brands, particularly Nissan and Chevrolet

oai_token-act-pair · gpt-3.5-turbo

references to automobile brands and models

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

psy

-0.73

 protocol

-0.69

aun

-0.67

 Protocol

-0.66

vere

-0.65

itu

-0.64

ienne

-0.64

 proxies

-0.64

 healing

-0.64

 wards

-0.64

POSITIVE LOGITS

 Nissan

2.92

 Toyota

2.84

 Mazda

2.79

 Volkswagen

2.77

 Chevrolet

2.77

BMW

2.69

 Porsche

2.64

 Chrysler

2.64

 Hyundai

2.61

 Volvo

2.54

Activations Density 0.046%

references to specific car brands, particularly Nissan and Chevrolet

references to automobile brands and models

No Comments

No Known Activations