INDEX

Explanations

the name "Harold" with varying levels of intensity

oai_token-act-pair · gpt-3.5-turbo

mentions of the name "Harold" across various contexts

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

eanor

-0.87

anwhile

-0.84

psey

-0.77

ocrats

-0.76

ocre

-0.75

ktop

-0.75

insula

-0.75

ramid

-0.74

*/(

-0.74

uyomi

-0.73

POSITIVE LOGITS

stein

0.93

 Harold

0.85

 McGee

0.85

fur

0.82

 Melvin

0.81

Kut

0.78

 Arthur

0.74

 Olson

0.74

 Rupert

0.73

ingly

0.73

Activations Density 0.022%

the name "Harold" with varying levels of intensity

mentions of the name "Harold" across various contexts

No Comments

No Known Activations