INDEX

Explanations

phrases indicating uncertainty or possibilities for future events

oai_token-act-pair · gpt-3.5-turbo

phrases indicating ongoing conditions or states that "remain" unchanged

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

catentry

-0.76

erity

-0.72

wcsstore

-0.69

 Surprise

-0.67

Mot

-0.65

BAT

-0.62

 Eclipse

-0.61

ATT

-0.60

elf

-0.59

dress

-0.58

POSITIVE LOGITS

 intact

0.89

 indefinitely

0.87

 haunt

0.83

 limbo

0.82

 untouched

0.81

 unchanged

0.79

 viable

0.78

lishes

0.78

 unaffected

0.76

 forever

0.74

Activations Density 0.144%

phrases indicating uncertainty or possibilities for future events

phrases indicating ongoing conditions or states that "remain" unchanged

No Comments

No Known Activations