INDEX

Explanations

the term "whatever" and its variations, indicating a focus on indifference or ambiguity

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

the word "whatever" in various contexts

oai_token-act-pair · gpt-3.5-turbo

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

agos

-0.85

ea

-0.83

enberg

-0.80

iard

-0.78

arie

-0.77

enburg

-0.76

ert

-0.75

press

-0.74

isse

-0.74

enthal

-0.74

POSITIVE LOGITS

 else

1.20

soever

1.13

 happens

1.10

 happened

1.07

 floats

0.96

 transpired

0.86

 semblance

0.84

THING

0.82

 kinds

0.76

 happ

0.75

Activations Density 0.031%

the term "whatever" and its variations, indicating a focus on indifference or ambiguity

the word "whatever" in various contexts

No Comments

No Known Activations