INDEX

Explanations

phrases indicating an outcome is not very probable

oai_token-act-pair · gpt-3.5-turbo

statements or phrases that express improbability or unlikelihood

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

ravings

-0.88

Ü

-0.80

lished

-0.77

ocked

-0.76

insula

-0.75

artney

-0.74

ricted

-0.74

CSS

-0.74

aeper

-0.74

aina

-0.74

POSITIVE LOGITS

icably

0.93

theless

0.87

bably

0.84

 unlikely

0.76

 necessarily

0.72

 unanim

0.71

 unanimous

0.71

 improbable

0.70

 coincidence

0.69

necess

0.69

Activations Density 0.019%

phrases indicating an outcome is not very probable

statements or phrases that express improbability or unlikelihood

No Comments

No Known Activations