INDEX

Explanations

words related to correct and incorrect judgments or evaluations

oai_token-act-pair · gpt-3.5-turbo

phrases or statements that reflect disbelief or discomfort

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

hement

-0.88

abouts

-0.86

nesday

-0.79

stad

-0.78

earch

-0.77

ortun

-0.72

isons

-0.72

 herself

-0.71

 intended

-0.70

osponsors

-0.70

POSITIVE LOGITS

ONSORED

0.87

Seriously

0.80

Unless

0.80

Liter

0.77

WHERE

0.77

HAHAHAHA

0.75

Okay

0.75

================================================================

0.73

Anyway

0.73

Domin

0.73

Activations Density 0.396%

words related to correct and incorrect judgments or evaluations

phrases or statements that reflect disbelief or discomfort

No Comments

No Known Activations