INDEX

Explanations

keywords related to statistics or attributes

oai_token-act-pair · gpt-3.5-turbo

statistical terms and references

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

 Sierra

-0.78

 Passage

-0.69

 Cascade

-0.69

 Clover

-0.67

bye

-0.67

 Whole

-0.65

 Horizons

-0.65

 tasting

-0.64

RIS

-0.64

¿½

-0.61

POSITIVE LOGITS

uer

0.96

 stat

0.95

ements

0.94

istically

0.92

ograph

0.91

ionics

0.91

ographs

0.91

iques

0.90

oids

0.89

uers

0.89

Activations Density 0.006%

keywords related to statistics or attributes

statistical terms and references

No Comments

No Known Activations