INDEX

Explanations

mentions of the word "Clark" with varying levels of activation

oai_token-act-pair · gpt-3.5-turbo

the name "Clark" and its various instances in the text

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

choes

-0.94

-0.93

urity

-0.89

ntil

-0.87

awaru

-0.81

orescence

-0.77

URI

-0.75

olitan

-0.73

exempt

-0.73

ItemTracker

-0.71

POSITIVE LOGITS

stown

1.02

 Clark

1.01

ston

0.94

 Kent

0.87

obyl

0.84

Clark

0.83

 Yards

0.82

 County

0.82

dale

0.79

anan

0.79

Activations Density 0.005%

mentions of the word "Clark" with varying levels of activation

the name "Clark" and its various instances in the text

No Comments

No Known Activations