INDEX

Explanations

names and locations, with a focus on a specific name "Hodgins" and locations like Wimbeldon and Hebbron

oai_token-act-pair · gpt-3.5-turbo

proper nouns, particularly names and locations

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

terday

-0.70

anwhile

-0.70

IZE

-0.69

 diving

-0.68

 backdrop

-0.67

 idol

-0.65

CLASSIFIED

-0.65

auts

-0.62

 Leone

-0.61

ãģį

-0.61

POSITIVE LOGITS

mination

1.04

ril

0.95

gers

0.94

rib

0.93

wig

0.92

rum

0.91

eless

0.91

rod

0.91

yssey

0.89

rans

0.88

Activations Density 0.034%

names and locations, with a focus on a specific name "Hodgins" and locations like Wimbeldon and Hebbron

proper nouns, particularly names and locations

No Comments

No Known Activations