INDEX

Explanations

scientific terms or references

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

drops

-0.72

zik

-0.72

edin

-0.69

tower

-0.69

pora

-0.69

ipop

-0.67

matched

-0.67

torn

-0.66

jar

-0.65

uble

-0.65

POSITIVE LOGITS

 fiction

1.02

 curiosity

1.01

 literacy

0.94

ĨĴ

0.91

 misconduct

0.90

 inquiry

0.89

 research

0.84

 endeavors

0.83

 Fiction

0.83

 hypotheses

0.83

Activations Density 0.022%

scientific terms or references

No Comments

No Known Activations