INDEX

Explanations

words related to logical or philosophical concepts such as paradoxes and dilemmas

oai_token-act-pair · gpt-3.5-turbo

concepts related to paradoxes and dilemmas

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

urate

-0.82

umin

-0.77

STD

-0.74

arate

-0.70

pecting

-0.68

uilt

-0.66

md

-0.66

È

-0.65

 unspecified

-0.65

Ò

-0.64

POSITIVE LOGITS

 dilemma

2.47

 paradox

2.20

 madness

1.80

 fiasco

1.79

 Paradox

1.67

 debacle

1.66

 frenzy

1.63

 predicament

1.58

 paradigm

1.49

cra

1.39

Activations Density 0.029%

words related to logical or philosophical concepts such as paradoxes and dilemmas

concepts related to paradoxes and dilemmas

No Comments

No Known Activations