INDEX

Explanations

sensory descriptions or visual observations related to seeing or observing

oai_token-act-pair · gpt-3.5-turbo

references to the concept of "sight" or visual experiences

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

 misunderstanding

-0.71

conn

-0.70

ulative

-0.66

 misunder

-0.65

oug

-0.64

currency

-0.64

 antidepressants

-0.64

isSpecialOrderable

-0.63

udeb

-0.62

iott

-0.61

POSITIVE LOGITS

seeing

1.60

orial

0.95

ight

0.89

lights

0.88

 Sight

0.82

enment

0.78

lessly

0.78

escription

0.76

ibility

0.76

ened

0.74

Activations Density 0.026%

sensory descriptions or visual observations related to seeing or observing

references to the concept of "sight" or visual experiences

No Comments

No Known Activations