INDEX

Explanations

the name "Joe" at various contexts

oai_token-act-pair · gpt-3.5-turbo

the name "Joe" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

 glim

-0.88

mble

-0.81

raints

-0.80

igators

-0.78

hips

-0.78

NESS

-0.76

lied

-0.74

igated

-0.73

dayName

-0.71

igator

-0.71

POSITIVE LOGITS

 Biden

0.95

zzi

0.87

 Arpaio

0.85

Rog

0.83

ppo

0.82

 Walsh

0.80

 Lieberman

0.76

Rao

0.75

 Camel

0.75

 Danger

0.73

Activations Density 0.008%

the name "Joe" at various contexts

the name "Joe" in various contexts

No Comments

No Known Activations