INDEX

Explanations

The neuron is looking for phrases related to physical flipping actions

oai_token-act-pair · gpt-3.5-turbo

phrases related to flipping or transforming objects or situations

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

EMENT

-0.83

ufact

-0.78

SPONSORED

-0.76

payer

-0.70

cliffe

-0.68

ICAN

-0.67

Rae

-0.65

 deductible

-0.64

 citiz

-0.63

 Shepherd

-0.63

POSITIVE LOGITS

ppy

0.84

ï¸ı

0.83

direction

0.76

nir

0.76

izont

0.69

coded

0.69

 punches

0.68

 blame

0.67

undo

0.67

oward

0.67

Activations Density 0.301%

The neuron is looking for phrases related to physical flipping actions

phrases related to flipping or transforming objects or situations

No Comments

No Known Activations