INDEX

Explanations

superlatives indicating the lowest degree or rank of something

oai_token-act-pair · gpt-3.5-turbo

references to the concept of "least" or minimal amounts

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

ses

-0.76

 Reborn

-0.72

gnu

-0.65

raltar

-0.64

 Dynamics

-0.64

xxxxxxxx

-0.63

xxxx

-0.62

anwhile

-0.62

kefeller

-0.61

ortium

-0.61

POSITIVE LOGITS

erest

0.82

 practicable

0.76

 intrusive

0.76

 toler

0.74

 conspicuous

0.72

 conceivable

0.71

 imaginable

0.70

 possible

0.69

 plausible

0.69

 restrictive

0.68

Activations Density 0.011%

superlatives indicating the lowest degree or rank of something

references to the concept of "least" or minimal amounts

No Comments

No Known Activations