INDEX

Explanations

phrases that indicate strong or firm positioning or support

oai_token-act-pair · gpt-3.5-turbo

expressions of certainty and strong assertions

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

vernment

-0.84

 Emin

-0.82

externalActionCode

-0.75

igraph

-0.75

ammy

-0.73

ebin

-0.73

nesota

-0.71

 tables

-0.70

EVA

-0.70

OTOS

-0.69

POSITIVE LOGITS

 entrenched

0.86

 rooted

0.86

ãĤ©

0.85

ens

0.85

lishes

0.85

 cement

0.84

 planted

0.83

 anchored

0.81

 grounded

0.76

 ingrained

0.76

Activations Density 0.014%

phrases that indicate strong or firm positioning or support

expressions of certainty and strong assertions

No Comments

No Known Activations