INDEX

Explanations

phrases related to representing or symbolizing something

oai_token-act-pair · gpt-3.5-turbo

phrases that express representation or significance

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

jet

-0.82

ithing

-0.80

seller

-0.75

strap

-0.73

page

-0.69

ny

-0.67

liner

-0.67

aired

-0.65

load

-0.64

raid

-0.64

POSITIVE LOGITS

ational

1.00

ATIVE

0.94

eering

0.84

atively

0.84

eers

0.75

¬¼

0.72

atives

0.70

ances

0.68

OUP

0.67

DonaldTrump

0.67

Activations Density 0.024%

phrases related to representing or symbolizing something

phrases that express representation or significance

No Comments

No Known Activations