INDEX

Explanations

phrases indicating a sequence of events

oai_token-act-pair · gpt-3.5-turbo

actions and events that are revealed or confirmed later in time

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

 yesterday

-0.83

 tomorrow

-0.76

 lately

-0.73

 today

-0.70

 currently

-0.69

 tonight

-0.67

 formerly

-0.66

 Thurs

-0.66

 constantly

-0.63

yet

-0.63

POSITIVE LOGITS

 regretted

0.77

 relent

0.68

 retract

0.67

iosyncr

0.66

 acknow

0.65

 regret

0.63

 forg

0.62

Reviewer

0.61

 apologise

0.61

gur

0.59

Activations Density 0.284%

phrases indicating a sequence of events

actions and events that are revealed or confirmed later in time

No Comments

No Known Activations