INDEX

Explanations

terms related to positive outcomes or rewards

oai_token-act-pair · gpt-3.5-turbo

terms associated with positive experiences and rewards

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

onne

-0.75

edia

-0.75

clerosis

-0.70

OPE

-0.70

sil

-0.68

efer

-0.68

behind

-0.67

owler

-0.65

peria

-0.65

olog

-0.64

POSITIVE LOGITS

tons

0.98

 corrid

0.94

ly

0.88

theless

0.83

 rewarding

0.79

LY

0.77

 inspirational

0.77

 conduc

0.75

emonic

0.73

itational

0.72

Activations Density 0.055%

terms related to positive outcomes or rewards

terms associated with positive experiences and rewards

No Comments

No Known Activations