INDEX

Explanations

instances of the word "again" occurring after a previous event

oai_token-act-pair · gpt-3.5-turbo

the phrase "once again" indicating repetition or continuity

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

regate

-0.71

-0.70

ocene

-0.70

adena

-0.66

owsky

-0.63

liest

-0.62

oslav

-0.62

uary

-0.61

 Institution

-0.60

ger

-0.59

POSITIVE LOGITS

ndra

0.73

forth

0.70

RGB

0.69

 proves

0.69

nces

0.68

 illustrating

0.67

mable

0.65

 cred

0.65

 conclud

0.64

 lapse

0.63

Activations Density 0.044%

instances of the word "again" occurring after a previous event

the phrase "once again" indicating repetition or continuity

No Comments

No Known Activations