INDEX

Explanations

proper nouns related to individuals named "Griffith" and "Fernando."

oai_token-act-pair · gpt-3.5-turbo

references to specific individuals' names

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

onyms

-0.88

enger

-0.88

ocular

-0.83

selves

-0.79

thal

-0.75

lished

-0.71

engers

-0.69

akable

-0.69

icist

-0.69

cule

-0.69

POSITIVE LOGITS

 Griffith

0.97

shire

0.73

 Palmer

0.67

 flowering

0.66

batch

0.63

 Inquisition

0.62

haw

0.60

rom

0.59

 Alonso

0.59

tree

0.59

Activations Density 0.038%

proper nouns related to individuals named "Griffith" and "Fernando."

references to specific individuals' names

No Comments

No Known Activations