INDEX

Explanations

lead researchers and investigators mentioned in research studies or investigations

oai_token-act-pair · gpt-3.5-turbo

references to academic research and its authors

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

 whine

-0.67

hhhh

-0.67

 stale

-0.66

nig

-0.64

hhh

-0.64

 cripp

-0.63

 hurting

-0.63

eeee

-0.63

 fireplace

-0.63

 flourished

-0.63

POSITIVE LOGITS

 research

0.89

study

0.85

earch

0.80

 advisory

0.78

 studies

0.78

 pioneering

0.78

 study

0.77

 groundbreaking

0.77

earcher

0.75

doctoral

0.74

Activations Density 0.305%

lead researchers and investigators mentioned in research studies or investigations

references to academic research and its authors

No Comments

No Known Activations