INDEX

Explanations

phrases indicating a conditional scenario

oai_token-act-pair · gpt-3.5-turbo

phrases indicating conditional necessity or requirements

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

paio

-0.81

jab

-0.67

apple

-0.65

lif

-0.65

car

-0.63

til

-0.62

dropping

-0.62

lda

-0.62

odore

-0.61

 Fidel

-0.61

POSITIVE LOGITS

 prompts

0.66

uum

0.63

 whatsoever

0.63

XD

0.62

entimes

0.61

 beforehand

0.59

 Were

0.59

 substant

0.59

 prompted

0.59

,...

0.59

Activations Density 0.054%

phrases indicating a conditional scenario

phrases indicating conditional necessity or requirements

No Comments

No Known Activations