INDEX

Explanations

the word "whatever" in different contexts

oai_token-act-pair · gpt-3.5-turbo

references to vague or undefined concepts

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

 Kamp

-0.69

restling

-0.65

nings

-0.64

kens

-0.61

 Auckland

-0.59

 Winnipeg

-0.59

waukee

-0.58

mberg

-0.58

bike

-0.58

tle

-0.58

POSITIVE LOGITS

 thereof

0.90

 assorted

0.84

 alike

0.83

whatever

0.82

 ensuing

0.78

 attendant

0.73

 accompanying

0.73

romeda

0.72

 consequ

0.72

il

0.71

Activations Density 0.268%

the word "whatever" in different contexts

references to vague or undefined concepts

No Comments

No Known Activations