INDEX

Explanations

time-related phrases indicating a specific period in the past

oai_token-act-pair · gpt-3.5-turbo

phrases related to time intervals, specifically recent years or decades

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

inki

-0.75

Fine

-0.71

aith

-0.68

cause

-0.65

oops

-0.64

angu

-0.61

because

-0.61

Tickets

-0.59

agra

-0.58

AB

-0.57

POSITIVE LOGITS

 there

0.81

we

0.80

 thereafter

0.71

 meanwhile

0.68

idents

0.67

 however

0.66

 moreover

0.65

 successive

0.62

SPONSORED

0.61

 psychologists

0.61

Activations Density 0.175%

time-related phrases indicating a specific period in the past

phrases related to time intervals, specifically recent years or decades

No Comments

No Known Activations