INDEX

Explanations

responses indicating discomfort or awkwardness in various contexts

oai_token-act-pair · gpt-3.5-turbo

words that indicate a description of something being "clunky" or awkward

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

awaru

-0.74

BALL

-0.70

chwitz

-0.69

Ô

-0.69

 Territory

-0.68

BIP

-0.68

GMT

-0.67

HAEL

-0.66

¥µ

-0.65

hyde

-0.64

POSITIVE LOGITS

amped

1.16

ipper

1.16

amps

1.15

ojure

1.12

ogged

1.04

ashing

1.03

avier

1.00

uster

0.99

ipp

0.98

aps

0.97

Activations Density 0.011%

responses indicating discomfort or awkwardness in various contexts

words that indicate a description of something being "clunky" or awkward

No Comments

No Known Activations