INDEX

Explanations

time-related expressions like "weeks ago" or "months back"

oai_token-act-pair · gpt-3.5-turbo

specific time references to past events

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

 substitutes

-0.66

itto

-0.65

 survives

-0.65

 controls

-0.62

urn

-0.62

 clen

-0.61

obe

-0.61

agos

-0.60

 dispers

-0.60

orting

-0.59

POSITIVE LOGITS

 rumors

0.75

ocument

0.74

stros

0.74

 commenter

0.72

lished

0.71

actionDate

0.70

reetings

0.67

ODUCT

0.66

 Wikileaks

0.65

published

0.65

Activations Density 0.273%

time-related expressions like "weeks ago" or "months back"

specific time references to past events

No Comments

No Known Activations