INDEX

Explanations

the word "nil."

oai_token-act-pair · gpt-3.5-turbo

terms related to the concept of "nil" and references to ozone or Mazda

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

sburg

-0.96

hou

-0.76

imation

-0.75

rets

-0.70

atters

-0.70

Tud

-0.70

roups

-0.69

esome

-0.68

asers

-0.67

prises

-0.67

POSITIVE LOGITS

uel

0.79

ilion

0.77

uncture

0.76

uppet

0.75

pire

0.72

clad

0.70

xon

0.70

ĸļ

0.68

senal

0.67

ournal

0.67

Activations Density 0.035%

the word "nil."

terms related to the concept of "nil" and references to ozone or Mazda

No Comments

No Known Activations