INDEX

Explanations

statements that are author or source attributions

oai_token-act-pair · gpt-3.5-turbo

sentences that contain periods or denote ends of thoughts

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 robber

-0.76

 criminal

-0.65

 decom

-0.64

 capsule

-0.63

heastern

-0.62

 crim

-0.62

icz

-0.62

 alloc

-0.62

 sandwich

-0.62

 weed

-0.61

POSITIVE LOGITS

obyl

0.93

txt

0.87

aughs

0.87

UGC

0.86

wav

0.85

gov

0.84

jpg

0.82

 Surely

0.81

 Unable

0.81

 Seems

0.81

Activations Density 0.372%

statements that are author or source attributions

sentences that contain periods or denote ends of thoughts

No Comments

No Known Activations