INDEX

Explanations

names or terms likely related to specific individuals or entities

oai_token-act-pair · gpt-3.5-turbo

different variations or references to the suffix "ien."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

ramid

-0.87

tm

-0.73

è¦ļéĨĴ

-0.70

atform

-0.70

rooms

-0.66

Discussion

-0.60

behind

-0.60

 subp

-0.59

 oats

-0.58

ritic

-0.57

POSITIVE LOGITS

vironment

1.14

cia

0.97

iewicz

0.96

hao

0.95

venue

0.95

kowski

0.91

ews

0.89

ci

0.88

stadt

0.84

za

0.83

Activations Density 0.020%

names or terms likely related to specific individuals or entities

different variations or references to the suffix "ien."

No Comments

No Known Activations