INDEX

Explanations

email addresses

oai_token-act-pair · gpt-3.5-turbo

occurrences of the domain "co.uk"

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

rhy

-0.64

 guiActiveUn

-0.62

ById

-0.62

 guiActive

-0.61

 shaving

-0.60

Bottom

-0.58

Pin

-0.57

 Advocate

-0.57

 caucuses

-0.56

 steroids

-0.56

POSITIVE LOGITS

uk

1.20

nz

1.12

za

0.96

jp

0.93

merce

0.89

ucl

0.87

gov

0.87

kr

0.86

dk

0.84

fecture

0.82

Activations Density 0.047%

email addresses

occurrences of the domain "co.uk"

No Comments

No Known Activations