INDEX

Explanations

the name "Joe" in various contexts

oai_token-act-pair · gpt-3.5-turbo

mentions of the name "Joe."

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

 glim

-0.88

mble

-0.82

hips

-0.79

raints

-0.78

igators

-0.77

NESS

-0.77

rights

-0.77

lied

-0.75

dayName

-0.75

liga

-0.73

POSITIVE LOGITS

 Biden

0.97

 Arpaio

0.85

Rog

0.85

ppo

0.84

zzi

0.83

 Walsh

0.80

 Lieberman

0.77

 Camel

0.77

 Danger

0.75

pport

0.73

Activations Density 0.011%

the name "Joe" in various contexts

mentions of the name "Joe."

No Comments

No Known Activations