INDEX

Explanations

proper nouns or names of people/places along with the verb "knows" indicating knowledge or awareness

oai_token-act-pair · gpt-3.5-turbo

statements of knowledge or awareness

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

phrine

-0.86

oples

-0.84

interstitial

-0.84

otion

-0.83

aredevil

-0.77

isco

-0.75

ItemTracker

-0.74

ermanent

-0.73

issions

-0.70

nesota

-0.70

POSITIVE LOGITS

ledged

1.19

ledge

1.08

lege

0.90

LED

0.86

how

0.82

ariat

0.78

ingly

0.75

nothing

0.73

 beforehand

0.73

 nothing

0.72

Activations Density 0.068%

proper nouns or names of people/places along with the verb "knows" indicating knowledge or awareness

statements of knowledge or awareness

No Comments

No Known Activations