INDEX

Explanations

terms related to processing or organizing information for efficiency

oai_token-act-pair · gpt-3.5-turbo

terms related to processes and activities associated with classification, preservation, and assessment

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

iott

-0.76

OPE

-0.75

isy

-0.75

ifi

-0.73

elo

-0.71

cci

-0.71

playing

-0.67

inel

-0.66

eday

-0.66

ãĥĩãĤ£

-0.65

POSITIVE LOGITS

redients

0.92

 spree

0.91

 ceremonies

0.88

 prowess

0.85

 techniques

0.84

HAM

0.83

tons

0.81

 ceremony

0.80

 rituals

0.79

 ability

0.77

Activations Density 0.320%

terms related to processing or organizing information for efficiency

terms related to processes and activities associated with classification, preservation, and assessment

No Comments

No Known Activations