INDEX

Explanations

short phrases related to actions or tasks

oai_token-act-pair · gpt-3.5-turbo

mentions of dates in the context of historical events or timelines

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

ãĤ´ãĥ³

-0.73

Cause

-0.67

ullivan

-0.66

DN

-0.66

zar

-0.65

FIX

-0.65

abil

-0.64

Topic

-0.64

grain

-0.64

arah

-0.62

POSITIVE LOGITS

 however

1.03

 meanwhile

0.92

 Mehran

0.72

 though

0.69

 Kahn

0.68

 although

0.66

 according

0.66

 when

0.66

 moreover

0.66

 Shank

0.62

Activations Density 0.183%

short phrases related to actions or tasks

mentions of dates in the context of historical events or timelines

No Comments

No Known Activations