INDEX

Explanations

mentions of mirrors

oai_token-act-pair · gpt-3.5-turbo

instances and variations of the word "mirror."

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

ensable

-0.77

--------------------------------------------------------

-0.76

stant

-0.74

cific

-0.72

ccess

-0.72

mercial

-0.70

aii

-0.69

reek

-0.69

roup

-0.69

RAFT

-0.69

POSITIVE LOGITS

ror

0.96

 Mirror

0.89

 mirror

0.82

 shine

0.81

ocular

0.80

 neuron

0.80

angelo

0.75

 reflection

0.75

 shards

0.72

pool

0.71

Activations Density 0.015%

mentions of mirrors

instances and variations of the word "mirror."

No Comments

No Known Activations