INDEX

Explanations

words or phrases related to suitability or alignment

oai_token-act-pair · gpt-3.5-turbo

phrases related to suitability or compatibility

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

dark

-0.70

ots

-0.68

awn

-0.67

inous

-0.66

gins

-0.65

ane

-0.64

uzzle

-0.63

rek

-0.63

ols

-0.63

rows

-0.61

POSITIVE LOGITS

fit

1.00

 suited

0.80

 bona

0.75

 adequately

0.71

 unbiased

0.71

 fitting

0.71

 suit

0.71

Fit

0.71

 snug

0.70

fit

0.70

Activations Density 0.278%

words or phrases related to suitability or alignment

phrases related to suitability or compatibility

No Comments

No Known Activations