INDEX

Explanations

events from a specific time frame, in this case, the previous week or month

oai_token-act-pair · gpt-3.5-turbo

instances of the word "Last" followed by a time period indicating recent events

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

Osc

-0.74

hare

-0.63

plain

-0.63

 Reborn

-0.62

aband

-0.60

cil

-0.59

ordes

-0.59

adle

-0.58

aunders

-0.58

entials

-0.58

POSITIVE LOGITS

 month

0.96

 week

0.94

 Month

0.86

 year

0.85

ime

0.82

ingly

0.77

ebin

0.77

 Gleaming

0.77

 ditch

0.76

 WEEK

0.76

Activations Density 0.034%

events from a specific time frame, in this case, the previous week or month

instances of the word "Last" followed by a time period indicating recent events

No Comments

No Known Activations