INDEX

Explanations

phrases related to summaries or synopses

oai_token-act-pair · gpt-3.5-turbo

mentions of summaries and summary reports

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

 charism

-0.83

gypt

-0.80

duct

-0.73

bees

-0.73

enne

-0.69

yss

-0.67

train

-0.66

 Mouth

-0.66

tan

-0.66

Wee

-0.65

POSITIVE LOGITS

 summ

0.77

 thereof

0.77

ured

0.74

 summar

0.72

 VIEW

0.72

 summarize

0.72

ARY

0.71

uration

0.69

ensus

0.69

 summary

0.69

Activations Density 0.055%

phrases related to summaries or synopses

mentions of summaries and summary reports

No Comments

No Known Activations