INDEX

Explanations

phrases describing a group or entity being compared, characterized, or evaluated in some way

oai_token-act-pair · gpt-3.5-turbo

phrases indicating quantity or classification descriptors

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

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

fax

-0.62

 veto

-0.61

alion

-0.61

atform

-0.61

 extradition

-0.61

icism

-0.60

formance

-0.60

ropri

-0.59

pipe

-0.59

plementation

-0.59

POSITIVE LOGITS

 selves

0.94

selves

0.90

 themselves

0.86

unts

0.82

 favorites

0.78

wolves

0.78

gh

0.78

united

0.76

 outnumbered

0.75

stars

0.75

Activations Density 0.281%

phrases describing a group or entity being compared, characterized, or evaluated in some way

phrases indicating quantity or classification descriptors

No Comments

No Known Activations